Research And Application Of Improved Long-Short-Term Memory Network In Chronic Disease Prediction

Chronic heart rate and blood pressure disease is a worldwide public health problem.In China,the incidence of heart failure is increasing year by year,but the effect of treatment is poor and the cost is high.About 35%of the people in China suffer from blood pressure diseases of varying degrees.These diseases and heart diseases form more serious complications,which seriously threaten the health of our residents.According to incomplete statistics,the number of patients with chronic heart rate and blood pressure diseases receiving antihypertensive and additional treatment in 2018 is over 32 million,and it is estimated that by 2020 it will exceed 40 million.At the end of 2018,the National Cardiovascular and Cerebrovascular Disease Foundation published a report in the Working Group on Chronic Cardiovascular and Cerebrovascular Diseases:The report shows that the prevalence of chronic heart rate and blood pressure diseases is high,and the mortality rate of chronic heart rate and blood pressure diseases is 10 to 30 times higher than that of non-blood pressure diseases.Based on the above,we can see that the problem of heart rate and blood pressure has increasingly become an important factor affecting the health of Chinese residents.Therefore,how to detect diseases early and timely is an urgent problem to be solved.Under this background,this paper first introduces the common prediction algorithms of chronic heart rate and blood pressure diseases,then gives a detailed introduction of the algorithms used in this paper,then carries out the prediction design of chronic heart rate and blood pressure diseases:data extraction,parameter selection and so on,and obtains the prediction results of the prediction model of chronic heart rate and blood pressure diseases based on LSTM,and finally finds out the prediction model based on LSTM.Deficiency of a predictive model for chronic heart rate and blood pressure disorders based on LSTM.In order to optimize the structure of LSTM,an improved LSTM model for predicting chronic heart rate and blood pressure disease was proposed.By analyzing the problems of LSTM in time series prediction,the fitting effect of the model was improved by adding feature extraction layer.By introducing the concept of feature extraction degree,the model can adapt to different application scenarios.The performance comparison shows the superiority of the proposed network structure based on LSTM.Finally,through the design and implementation of simulation and prediction program,the practical application of the improved algorithm is simulated,and the practicability of the improved algorithm is proved.